Geographic and temporal trends in antimicrobial nonsusceptibility in Streptococcus pneumoniae in the post-vaccine era in the United States.

We examined whether observed increases in antibiotic nonsusceptible nonvaccine serotypes after introduction of pneumococcal conjugate vaccine in the United States in 2000 were driven primarily by vaccine or antibiotic use. Using active surveillance data, we evaluated geographic and temporal differences in serotype distribution and within-serotype differences during 2000-2009. We compared nonsusceptibility to penicillin and erythromycin by geography after standardizing differences across time, place, and serotype by regressing standardized versus crude proportions. A regression slope (RS) approaching zero indicates greater importance of the standardizing factor. Through 2000-2006, geographic differences in nonsusceptibility were better explained by within-serotype prevalence of nonsusceptibility (RS 0.32, 95% confidence interval [CI], .08-.55 for penicillin) than by geographic differences in serotype distribution (RS 0.71, 95% CI, .44-.97). From 2007-2009, serotype distribution differences became more important for penicillin (within-serotype RS 0.52, 95% CI, .11-.93; serotype distribution RS 0.57, 95% CI, .14-1.0). Differential nonsusceptibility, within individual serotypes, accounts for most geographic variation in nonsusceptibility, suggesting selective pressure from antibiotic use, rather than differences in serotype distribution, mainly determines nonsusceptibility patterns. Recent trends suggest geographic differences in serotype distribution may be affecting the prevalence of nonsusceptibility, possibly due to decreases in the number of nonsusceptible serotypes.